- Email: [email protected]
Up-regulation of Fas ligand and down-regulation of Fas expression in oral carcinogenesis
Oral Oncology 35 (1999) 548±553
Up-regulation of Fas ligand and down-regulation of Fas expression in oral carcinogenesis Qianming Chen a, b,*, L.P. Samaranayake a, Xin Zhen b, Gang Luo b, Minhai Nie b, Bingqi Li b a
Oral Bio-Science Laboratories, Faculty of Dentistry, The University of Hong Kong, Hong Kong College of Stomatology, West China University of Medical Sciences, People's Republic of China
b
Received 22 January 1999; accepted 11 February 1999
Abstract An important molecule involved in delivering the death signal that initiates apoptosis is called Fas, or Apo-1, which sits on the cell surface. When another molecule called the Fas ligand (FasL) binds to it, Fas triggers a series of events inside the cell that leads to apoptosis. In order to investigate the mechanism of immune escape and the expression of Fas and FasL in oral premalignant lesions (OPLs) and oral squamous cell carcinomas (OSCCs), a total of 64 samples were evaluated by an immunohistochemical method using a labelled streptavidin±biotin assay. These samples comprised nine hyperkeratotic and 24 oral premalignant lesions (nine of mild, moderate, and six of severe dysplastic lesions), and 24 OSCCs, together with seven healthy controls. The results demonstrated that the majority of invasive OSCCs showed down-regulation of Fas expression but up-regulation of FasL expression. These phenomena were also detected in OPLs. The results indicate that the expression of Fas and FasL is involved in oral carcinogenesis and this may be a mechanism by which the cancer cells evade the host immune assault. Perhaps, in future, Fas/FasL system may be used as a prognostic biomarker in predicting the behavior of oral premalignant lesions. # 1999 Elsevier Science Ltd. All rights reserved. Keywords: Fas; FasL; Oral; Premalignant lesions; Squamous cell carcinoma
1. Introduction An important molecule involved in delivering the death signal that initiates apoptosis (programmed cell death) is called Fas, or Apo-1/CD95 [1,2]. When another molecule called the Fas ligand (FasL) binds to it, Fas triggers a series of events inside the cell that lead to apoptosis [1,2]. Fas is a member of the receptor family of cell surface proteins, and FasL belongs to the tumor necrosis factor superfamily of membrane and secreted proteins [1,2]. In addition to expression on the cells of the immune system such as T cells, B cells, and natural killer cells, Fas is also expressed in adult tissues, including liver, ovary, thymus, heart, and skin [1±5]. FasL expression was ®rst considered to be restricted to the immune system, including activated T cells, B cells, and natural killer cells designed to kill other cells * Corresponding author. Oral Bio-Science Laboratories, Faculty of Dentistry, The University of Hong Kong, Hong Kong. Tel.: +8522859-0481; fax: +852-2547-6133. E-mail address: [email protected] (Qianming Chen)
[6±8]. However, the identi®cation of FasL expression by cells in immune-privileged sites, such as stroma cells of the eye and Sertoli cells of the testis, suggests that activated T cells bearing Fas would encounter cells expressing FasL and hence receive a death signal, obviating an immune response [1]. These observations indicate that certain non-immune cells in immune-privileged sites can use Fas/FasL system against T cells themselves [1]. Recently, induction or up-regulation of functional FasL expression has been detected in several human neoplasms, such as oesophagus, lung, glia, colon, pancreas, etc., suggesting that tumour cells themselves might be able to defend against attack from T cells or other eector cells of the immune system by expressing FasL [9±16]. In this manner, tumorigenic cells appear to successfully escape the monitoring of the immune system just as do the immune-privileged cells [10]. Although there is some information on Fas expression in both healthy and diseased oral tissues, similar data on FasL expression are scarce. Yoshioka et al. [17] raised a polyclonal antibody against a synthetic polypeptide selected from the extracellular region of the
1368-8375/99/$ - see front matter # 1999 Elsevier Science Ltd. All rights reserved. PII: S1368-8375(99)00029-9
Qianming Chen et al. / Oral Oncology 35 (1999) 548±553
human Fas antigen and employed this for evaluation of the latter antigen in human gingivae. They demonstrated Fas expression in human gingivae which in turn was related to the physiological turnover of the epithelium. Further, they used this antibody to detect Fas expression in odontogenic keratocysts, leukoplakia, lichen planus, and squamous cell carcinoma (SCC) [18], and showed characteristic patterns of Fas antigen distribution for each disease entity with some degree of overlap. Recently, Dekker et al. [19] reported no dierences in Fas and FasL expression in oral lichen planus and healthy controls, while Chrysomali et al. [20] found minor dierences in staining patterns in oral epithelial tissues from erythema multiforme patients and controls. However, there exist contradictory results on Fas expression in oral carcinogenesis. Muraki et al. [18] reported that the expression of Fas decreased in epithelia derived from oral squamous cell carcinomas (OSCCs), while Sundelin et al. [21] found increased Fas in lingual SCC with no expression in healthy oral mucosa. There is, thus far, a single paper relating to FasL expression in oral carcinogenesis by Lora et al. [22]. They reported that FasL expression correlated negatively with degree of dierentiation and apoptosis in OSCCs after studying eight samples of normal oral mucosa and 19 OSCCs using immunohistochemistry and the TUNEL methods [22]. We, therefore, studied the expression of Fas and FasL in various stages of oral carcinogenesis using a previously described immunohistochemical method [23]. 2. Materials and methods 2.1. Samples A total of 64 samples were studied; 57 samples were obtained from patients attending the Stomatological Hospital, West China University of Medical Sciences, Chengdu, Sichuan, People's Republic of China, during 1994±97. They were all archival specimens graded according to the criteria of an international collaborative group on oral white lesions and the World Health Organization on oral cancers [24,25]. These included nine cases with hyperkeratosis, 24 oral premalignant lesions (OPLs, nine of mild, moderate, and six of severe dysplastic lesions), 24 OSCCs. None of the patients had radiotherapy or chemotherapy or other interventional palliative or therapeutic measures prior to sampling. Control samples of oral mucosa (5 mm diameter) were obtained from seven individuals undergoing plastic surgery, all of whom gave informed consent for the use of their tissue in the experimental procedures. The specimens were obtained and blocked, ®xed with 10% buered formalin and embedded in paran
549
using conventional histopathological techniques. Sections (3-mm) were prepared and evaluated after staining with haematoxylin and eosin. Their histopathologic grades were recon®rmed according to the standard criteria [24,25] by two experienced histopathologists, in a double-blind manner. 2.2. Antibodies The Fas and FasL rabbit polyclonal antibody was commercially purchased from Santa Cruz Biotechnology, Inc., CA, USA. According to the manufacturer, the Fas antibody was raised in rabbits immunised with a peptide sequence mapping the carboxy terminus of the Fas precursor of human origin, and was speci®c for the latter. The FasL antibody was also raised in rabbits immunised with a peptide sequence mapping the amino terminus of FasL of rat origin, and reacted with FasL from mouse, rat and humans. 2.3. Staining procedure Immunohistochemistry for detecting both Fas and FasL was performed as described previously, with a slight modi®cation, using a labelled streptavidin±biotin (LSAB) immunohistochemistry assay [23]. In brief, 3mm-thick sections of paran-embedded tissues were cut, mounted onto slides coated with 5% APES (3aminopropyltriethoxysilane, Sigma), dried overnight at 56 C, dewaxed in xylene and rehydrated through descending graded alcohols to phosphate-buered saline (PBS, pH 7.4). For antigen retrieval, slides were immersed in 10 mM sodium citrate buer (pH 6.0) and boiled twice for 5 min in a microwave oven (800 W). The liquid level was adjusted with addition of distilled water after the ®rst 5 min. Slides were then allowed to cool down in the citrate buer and washed thoroughly in three changes of PBS. In order to inhibit endogenous peroxidase activity, the slides were treated for 10 min with 3% (v/v) H2O2, 18% methanol (v/v) in PBS. Then, they were placed in a humidi®ed chamber and covered with 10% normal porcine serum (Dako, Denmark) for 10 min at 37 C to block non-speci®c proteinbinding sites. After decanting the excess serum, slides were incubated with primary antibody (diluted 1:400 for Fas and 1:100 for FasL) for 60 min, at 37 C for Fas and, overnight, at 4 C for FasL. Slides were subsequently incubated with prediluted biotinylated antibody against rabbit immunoglobulins (Dako), and streptavidin± horseradish peroxidase conjugate for 30 min at 37 C. After washing, peroxidase activity was detected using 3,30 -diaminobenzidine (Dako) as chromogen with H2O2 as substrate. The sections were counterstained with Harris haematoxylin, dehydrated, cleared in xylene, and mounted in DePeX (DaKo).
550
Qianming Chen et al. / Oral Oncology 35 (1999) 548±553
2.4. Interpretation of results The cells in the test specimens, which demonstrated granular brown staining, were considered as positive. The expression of Fas and FasL was estimated semiquantitatively as follows: ÿ, no expression; + (mild), <5% positive cells; ++ (moderate), 6±25% positive cells; +++ (intense), 26±50% positive cells; or ++++ (very intense), >50% positive cells [16]. To minimize inter-examiner variation, each slide was evaluated by two experienced histopathologists and the specimens with more than 10% deviation in the recorded staining intensity index were reassessed using a multi-headed microscope to achieve consensus. 2.5. Controls Positive controls were included in each assay and comprised formalin-®xed, paran-embedded sections of normal human lymph nodes. Blank controls were fabricated for each specimen by the omission of the primary antibody, which was replaced with PBS. Substitution of the primary antibody with antibody-free serum was also used as a replacement control. 2.6. Statistical analysis Kruskal±Wallis test was utilized to evaluate the differences in the staining intensity of Fas or FasL in each group. Dierences at 5% level were considered statistically signi®cant. All statistical evaluations were performed using the SPSS version 6.0 package (SPSS Inc., Chicago, USA).
Table 1 Expression of Fas in oral carcinogenesisa Staining
N
HK MiD MoD SD WOSCC MOSCC POSCC
++++ 0 (Very severe)
0
0
0
0
0
0
0
+++ (Severe)
1
2
0
0
0
0
0
0
++ (Moderate)
5
6
4
2
1
0
0
0
+ (Mild)
1
1
4
7
3
6
2
0
ÿ (Negative)
0
0
1
0
2
6
6
4
Total number 7 of specimens
9
9
9
6
12
8
4
a Staining: Fas and FasL expression were semi-quantitatively evaluated as follows: ÿ, no expression; +, <5% positive cells; ++, 5± 25%; +++, 26±50% positive cells; ++++, >50% positive cells. N, normal controls; HK, hyperkeratosis; MiD, mild dysplasia; MoD, moderate dysplasia; SD, severe dysplasia; WOSCC, well-dierentiated oral squamous cell carcinomas (OSCCs); MOSCC, moderately dierentiated OSCCs; POSCC, poorly dierentiated OSCCs.
tended to be either equivalent to or greater than that of the normal mucosa (Fig. 1b). In the invasive OSCC group, the degree of expression of Fas antigen decreased signi®cantly when compared with the controls ( p<0.05) and seemed to relate well to the degree of tumour dierentiation. For example, in the 12 well-dierentiated tumors, six samples demonstrated a mild degree of antigen-positive cells (<5%, according to the grading system) particularly in cell nests (Fig. 1c). No Fas-positive cells were observed in the epithelial cells in poorly dierentiated OSCCs.
3. Results
3.2. FasL staining in oral epithelia
3.1. Fas staining in oral epithelia
Generally, FasL expression was mainly seen in the cytoplasm, with little or no expression on the cell membrane. The overall results of FasL expression in healthy and diseased tissues are shown in Table 2. FasL could be detected in all normal oral mucosal tissues, particularly localized in the basal cells, with a weak reaction in the prickle cell layers (Fig. 2a). In hyperkeratotic tissues and OPLs, the staining pattern of FasL was similar to normal controls, although the number of FasL-expressing cells was signi®cantly higher ( p<0.05) (Fig. 2b). All 24 specimens from OSCCs demonstrated increased staining intensity as well as cell numbers with FasL ( p<0.05; Fig. 2c) when compared with both normal controls and OPLs.
The Fas antigen appeared as granular material essentially on the cell membrane of the epithelium, although some positive reaction was visualised in the cytoplasm. The results of Fas expression during various stages of oral carcinogenesis are shown in Table 1. Immunohistochemically, Fas was expressed in all seven normal oral mucosal specimens. This was clearly evident in more dierentiated layers such as the granular and the upper part of the prickle layer, but not in the basal cell layers (Fig. 1a). Varying degrees of Fas expression were seen in the diseased tissues. The staining pattern of all nine hyperkeratotic lesions was similar to the normal oral mucosa. Quantitatively, the number of positive staining cells in the three OPL groups appeared to be lower when compared with the normal controls. This was particularly true in the moderately and severely dysplastic tissues ( p<0.05). However, the intensity of staining
3.3. Fas and FasL in the lymphocytic in®ltrate Both Fas and FasL could be detected in the lymphocytic in®ltrate of the lamina propria of the normal
Qianming Chen et al. / Oral Oncology 35 (1999) 548±553
551
Table 2 Expression of FasL in oral carcinogenesisa Staining
N
HK MiD MoD SD WOSCC MOSCC POSCC
++++ 0 (Very severe)
0
0
1
3
8
4
4
+++ (Severe)
0
1
4
6
1
3
3
0
++ (Moderate)
1
5
1
2
2
1
1
0
+ (Mild)
6
2
3
0
0
0
0
0
ÿ (Negative)
0
1
1
0
0
0
0
0
Total number 7 of specimens
9
9
9
6
12
8
4
a Staining: Fas or FasL expression were semi-quantitatively evaluated as follows: ÿ, no expression; +, <5% positive cells; ++, 5± 25%; +++, 26±50% positive cells; ++++, >50% positive cells. N, normal controls; HK, hyperkeratosis; MiD, mild dysplasia; MoD, moderate dysplasia; SD, severe dysplasia; WOSCC, well-dierentiated oral squamous cell carcinomas (OSCCs); MOSCC, moderately dierentiated OSCCs; POSCC, poorly dierentiated OSCCs.
the normal mucosa. However, the number of antigenbearing cells tended to increase proportionally with the degree of lymphocytic in®ltrate. 3.4. The controls In each case, positive controls (i.e. human lymph node tissues) for both Fas and FasL, were consistently positive, whereas negative controls (sections without primary antibody and sham rabbit serum for Fas and FasL) were consistently negative. 4. Discussion
Fig. 1. Fas immunohistochemistry in oral carcinogenesis. (a) Fas expression in normal oral mucosa, predominantly in the cells of the granular layer and the upper part of the prickle layer; no Fas antigen was detected in the basal layers but some reaction was noted in the in®ltrating lymphocytes (arrow). (b) Some evidence of Fas staining on cell surface in moderate dysplasia. (c) A well-dierentiated tumour, where the Fas reaction appeared mainly in cell nests.
mucosa (Figs. 1a and 2a) as well as OPLs (Fig. 2a and b), and in all the tumours (Fig. 2c). The degree of expression of both these antigens in the lymphocytic in®ltrate of OPLs and OSCCs was similar to that of
Our results clearly indicate that Fas antigen is expressed to some extent in the well-dierentiated layers of the normal oral mucosa as well as in cells of welldierentiated SCCs. The latter was present mostly in the centre of tumour nests. These ®ndings seem to suggest that Fas expression is related to the degree of dierentiation of the squamous epithelium, not only in normal but also in carcinomatous tissues. These results concur with those of Yoshioka et al. [17] and Gratsa et al. [16] who reported down-regulation of Fas expression in oral mucosa and oesophageal mucosa, respectively. However, Sundelin et al. [21] could not detect Fas expression in normal oral mucosa. We also clearly demonstrate Fas positivity particularly in cell membranes, a ®nding consistent with previous observations that Fas is a type II membrane protein [1,2] where the carboxy terminus is membranic.
552
Qianming Chen et al. / Oral Oncology 35 (1999) 548±553
Fig. 2. FasL immunohistochemistry in oral carcinogenesis. (a). Fas L staining cells were detected in the normal oral mucosa, mainly in the basal cells, with a weak reaction in the prickle cell layer. The expression of FasL was also seen in the in®ltrating lymphocytes in the lamina propria (arrow). (b) Moderate dysplasia: the cell number and the intensity of FasL expression is increased compared with normal controls. (c) Dysplastic epithelium from an oral squamous cell carcinoma (OSCC) demonstrating increased FasL staining intensity compared with normal controls as well as the oral premalignant lesions, with simultaneous FasL expression in in®ltrating lymphocytes (arrow).
The Fas antibody, which we used in the study, was raised with a peptide sequence mapping at the carboxy terminus of the Fas. Fas expression was signi®cantly down-regulated in the oral premalignant lesions. Indeed, in the majority of OSCCs investigated such expression was completely lost or limited to a small population of neoplastic cells. This suggests that down-regulation of Fas expression may be an early event in oral carcinogenesis in an attempt to escape immune surveillance and destruction by activated cytotoxic T cells [9±15]. Similar ®ndings have been reported by Strand et al. [9], Grate et al. [16], Yoshioka et al. [17] in hepatocellular, oesophageal, and oral cancers, respectively. However, these reports are inconsistent with the ®ndings of Sundelin et al. [21], who found increased Fas expression in SCC of the tongue. We also demonstrate here that FasL is constitutively expressed in the normal oral mucosa, predominantly in the basal and suprabasal layers. The biological signi®cance of FasL expression in such a graded fashion in normal oral epithelium needs to be further investigated. As there is a degree of lymphocytic in®ltrate in normal oral mucosa [26], it is temping to speculate that FasL expression may contribute to the protection of basal cells from apoptotic cell death induced by lymphocytes that are present in the region. FasL positivity was detected mostly in the cellular cytoplasm both in the normal tissues and the diseased epithelia, thus agreeing with previous studies [1,2]. This result is not surprising as the amino terminus in this family of proteins is cytoplasmic, and the FasL antibody we used in this study was raised in rabbits immunised with a peptide sequence mapping at the amino terminus. Our results imply that up-regulation of FasL may be common in OSCCs as this phenomenon was readily detectable in oral premalignant lesions. In general, the degree of FasL expression concurred with the degrees of dysplasia, suggesting this to be an early event in the natural history of oral carcinogenesis. This is consistent with the recent report by Lora et al. [22] who correlated the immunohistochemical expression of FasL with the dierentiation of OSCCs and found that FasL expression was highly expressed in poorly dierentiated lesions and in some moderately dierentiated lesions. We have also demonstrated that the lymphocytic in®ltrate in OSCCs expressed both Fas and FasL whereas in a previous study we reported that the number of the in®ltrating lymphocytes signi®cantly decreased in such lesions [26]. These may be related events. Recent reports on hepatocellular carcinomas [15], melanomas [10], lung carcinomas [11], glioblastomas [12,13], colon carcinomas [14], and oesophageal carcinomas [16] clearly suggest that the FasL expressed in carcinoma cells is functional. Such functional FasL in tumour cells is likely to defend the cells against attack from the cytotoxic T cells or
Qianming Chen et al. / Oral Oncology 35 (1999) 548±553
other eector cells of the immune system by triggering an apoptotic response in lymphocytes [10]. Therefore, we propose that up-regulation of FasL is a defensive strategy of neoplastic cells of the oral epithelium attempting to escape immune rejection. In summary, constitutive expression of Fas appears to correlate with the degree of dierentiation of the normal oral mucosa. The majority of invasive OSCCs demonstrated down-regulation of Fas-expressing cells and simultaneous in®ltration by FasL-expressing lymphocytes, suggesting this to be a mechanism by which these cells evade Fas-mediated cell death. Furthermore, the majority of invasive OSCCs show up-regulation of FasL when compared with the premalignant lesions, implying that tumour cells may counteract the immune system by killing Fas-expressing cytotoxic T cells. Thus, upregulation of FasL and down-regulation of Fas appear to be simultaneous events in oral carcinogenesis. These results, taken together, indicate that the Fas/FasL system may serve as a useful clinical tool in the future as a prognostic biomarker in oral premalignant lesions. Acknowledgements This work was supported in part by a grant from the National Nature Science Foundation in China (No. 39300149), a grant from the Ministry of Public Health in China (No. 9504125), and a grant from the Committee on Research and Conference Grants of the University of Hong Kong, Hong Kong (No. 10201937/ 30713/08011/301/01). References [1] Nagato S. Apoptosis by death factor. Cell 1997;88:355±65. [2] Nagato S, Golstein P. The Fas death factor. Science 1995; 267:1449±56. [3] Leithauser F, Dhein J, Mechtersheimer G, et al. Constitutive and induced expression of APO-1, a new member of the nerve growth factor/tumor necrosis factor receptor superfamily, in normal and neoplastic cells. Laboratory Investigation 1993;69:415±29. [4] Watanabe-Fukunaga R, Brannan CI, et al. The cDNA structure, expression and chromosomal assignment of the mouse Fas antigen. Journal of Immunology 1992;148:1274±9. [5] Frech LE, Hahne M, Viard I, et al. Fas and Fas ligand in embryos and adult mice: ligand expression in several immune-privileged tissues and coexpression in adult tissues characterized by apoptotic cell turnover. Journal of Cell Biology 1996;133:335±43. [6] Suda T, Takahashi T, Golstein P, Nagata S. Molecular cloning and expression of the Fas ligand, a novel member of the tumor necrosis factor family. Cell 1993;75:1169±78. [7] Hahne M, Renno T, Schroeter M, et al. Activated B cells express functional fas ligand. European Journal of Immunology 1996; 26:721±4. [8] Loughran TPJ. Clonal diseases of large granular lymphocytes. Blood 1993;82:1±14.
553
[9] Strand S, Hofmann WJ, Hug H, et al. Lymphocyte apoptosis induced by CD95 (APO-1/Fas) ligand-expressing tumor cells: a mechanism of immune evasion? Nature Medicine 1996;2: 1361±6. [10] Hahne M, Rimoldi D, Schroter M, et al. Melanoma cell expression of Fas (Apo-1/CD95) ligand: implications for tumor immune escape. Science 1996;274:1363±6. [11] Niehans GA, Brunner T, Frizelle SP, et al. Human lung carcinomas express Fas ligand. Cancer Research 1997;57:1007±12. [12] Saas P, Walker PR, Hahne M, et al. Fas ligand expression by astrocytoma in vivo: maintaining immune privilege in the brain? Journal of Clinical Investigation 1997;99:1173±8. [13] Gratas C, Tohma Y, Van Meir EG, et al. Fas ligand expression in glioblastoma cell lines and primary astrocytic brain tumors. Brain Pathology 1997;7:863±9. [14] O'Connell J, O'Sullivan GC, Collins JK, Shanahan F. The Fascounterattack: Fas-mediated T cell killing by colon cancer cells expressing Fas ligand. Journal of Experimental Medicine 1996; 184:1075±82. [15] Shiraki K, Tsuji N, Shioda T, Isselbacher KJ, Takahashi H. Expression of Fas ligand in liver metastases of human colonic adenocarcinomas. Proceedings of the National Academy of Science, USA 1997;94:6420±5. [16] Gratas C, Tohma Y, Barnas C, et al. Up-regulation of Fas (APO1/CD95) ligand and down-regulation of Fas expression in human esophageal cancer. Cancer Research 1998;58:2057±62. [17] Yoshioka C, Muraki Y, Fukuda J, Haneji T, Kobayashi N. Identi®cation of the Fas antigen in human gingiva. Journal of Dental Research 1996;75:1353±7. [18] Muraki Y, Yoshioka C, Fukuda J, Haneji T, Kobayashi N. Immunohistochemical detection of Fas antigen in oral epithelia. Journal of Oral Pathology and Medicine 1997;26:57±62. [19] Dekker NP, Lozada-Nur F, Lagenaur LA, MacPhail LA, Bloom CY, Regezi JA. Apoptosis-associated markers in oral lichen planus. Journal Oral Pathology and Medicine 1997; 26:170±5. [20] Chrysomali E, Dekker NP, Regezi J. Apoptosis in oral erythema multiforme. Oral Surgery Oral Medine Oral Pathology Oral Radiology Endodontics 1997;83:272±80. [21] Sundelin K, Jadner M, Norberg-Spaak L, Davidsson A, Hellquist HB. Metallothionein and Fas (CD95) are expressed in squamous cell carcinoma of the tongue. European Journal of Cancer 1997;33:1860±4. [22] Lora LL, Vintermyr OK, Johannessen AC, Liavaag PG, Jonsson R. Suppression of Fas receptor and negative correlation of Fas ligand with dierentiation and apoptosis in oral squamous cell carcinoma. Journal of Oral Pathology and Medicine 1999;28: 82±7. [23] Chen Q, Luo G, Li B, Samaranayake LP. Expression of p16 and cdk4 in oral premalignant lesions and oral squamous cell carcinomas: a semi-quantitative immunohistochemical study. Journal of Oral Pathology and Medicine 1999;28:158±64. [24] Axell T, Pindborg JJ, Smith CJ, Waal Van Der I. Oral white lesions with special reference to precancerous and tobacco-related lesions: conclusions of an international symposium held in Uppsala, Sweden, May 18±21, 1994. Journal of Oral Pathology and Medicine 1996;25:49±54. [25] Pindborg JJ, Reichart PA, Smith CJ, Waal van der I. Histological Typing of Cancer and precancer of the Oral Mucosa, 2nd Edition. World Health Organization, Springer, 1997, p. 11±13. [26] Nie M, Chen Q, Li B, Zhen G, Luo G, Zhen X. A quantitatively pathological study on the lymphocytic in®ltrate in oral carcinogenesis. Clinical Journal of Stomatology 1997;13:213±5 (in Chinese).
Up-regulation of Fas ligand and down-regulation of Fas expression in oral carcinogenesis Qianming Chen a, b,*, L.P. Samaranayake a, Xin Zhen b, Gang Luo b, Minhai Nie b, Bingqi Li b a
Oral Bio-Science Laboratories, Faculty of Dentistry, The University of Hong Kong, Hong Kong College of Stomatology, West China University of Medical Sciences, People's Republic of China
b
Received 22 January 1999; accepted 11 February 1999
Abstract An important molecule involved in delivering the death signal that initiates apoptosis is called Fas, or Apo-1, which sits on the cell surface. When another molecule called the Fas ligand (FasL) binds to it, Fas triggers a series of events inside the cell that leads to apoptosis. In order to investigate the mechanism of immune escape and the expression of Fas and FasL in oral premalignant lesions (OPLs) and oral squamous cell carcinomas (OSCCs), a total of 64 samples were evaluated by an immunohistochemical method using a labelled streptavidin±biotin assay. These samples comprised nine hyperkeratotic and 24 oral premalignant lesions (nine of mild, moderate, and six of severe dysplastic lesions), and 24 OSCCs, together with seven healthy controls. The results demonstrated that the majority of invasive OSCCs showed down-regulation of Fas expression but up-regulation of FasL expression. These phenomena were also detected in OPLs. The results indicate that the expression of Fas and FasL is involved in oral carcinogenesis and this may be a mechanism by which the cancer cells evade the host immune assault. Perhaps, in future, Fas/FasL system may be used as a prognostic biomarker in predicting the behavior of oral premalignant lesions. # 1999 Elsevier Science Ltd. All rights reserved. Keywords: Fas; FasL; Oral; Premalignant lesions; Squamous cell carcinoma
1. Introduction An important molecule involved in delivering the death signal that initiates apoptosis (programmed cell death) is called Fas, or Apo-1/CD95 [1,2]. When another molecule called the Fas ligand (FasL) binds to it, Fas triggers a series of events inside the cell that lead to apoptosis [1,2]. Fas is a member of the receptor family of cell surface proteins, and FasL belongs to the tumor necrosis factor superfamily of membrane and secreted proteins [1,2]. In addition to expression on the cells of the immune system such as T cells, B cells, and natural killer cells, Fas is also expressed in adult tissues, including liver, ovary, thymus, heart, and skin [1±5]. FasL expression was ®rst considered to be restricted to the immune system, including activated T cells, B cells, and natural killer cells designed to kill other cells * Corresponding author. Oral Bio-Science Laboratories, Faculty of Dentistry, The University of Hong Kong, Hong Kong. Tel.: +8522859-0481; fax: +852-2547-6133. E-mail address: [email protected] (Qianming Chen)
[6±8]. However, the identi®cation of FasL expression by cells in immune-privileged sites, such as stroma cells of the eye and Sertoli cells of the testis, suggests that activated T cells bearing Fas would encounter cells expressing FasL and hence receive a death signal, obviating an immune response [1]. These observations indicate that certain non-immune cells in immune-privileged sites can use Fas/FasL system against T cells themselves [1]. Recently, induction or up-regulation of functional FasL expression has been detected in several human neoplasms, such as oesophagus, lung, glia, colon, pancreas, etc., suggesting that tumour cells themselves might be able to defend against attack from T cells or other eector cells of the immune system by expressing FasL [9±16]. In this manner, tumorigenic cells appear to successfully escape the monitoring of the immune system just as do the immune-privileged cells [10]. Although there is some information on Fas expression in both healthy and diseased oral tissues, similar data on FasL expression are scarce. Yoshioka et al. [17] raised a polyclonal antibody against a synthetic polypeptide selected from the extracellular region of the
1368-8375/99/$ - see front matter # 1999 Elsevier Science Ltd. All rights reserved. PII: S1368-8375(99)00029-9
Qianming Chen et al. / Oral Oncology 35 (1999) 548±553
human Fas antigen and employed this for evaluation of the latter antigen in human gingivae. They demonstrated Fas expression in human gingivae which in turn was related to the physiological turnover of the epithelium. Further, they used this antibody to detect Fas expression in odontogenic keratocysts, leukoplakia, lichen planus, and squamous cell carcinoma (SCC) [18], and showed characteristic patterns of Fas antigen distribution for each disease entity with some degree of overlap. Recently, Dekker et al. [19] reported no dierences in Fas and FasL expression in oral lichen planus and healthy controls, while Chrysomali et al. [20] found minor dierences in staining patterns in oral epithelial tissues from erythema multiforme patients and controls. However, there exist contradictory results on Fas expression in oral carcinogenesis. Muraki et al. [18] reported that the expression of Fas decreased in epithelia derived from oral squamous cell carcinomas (OSCCs), while Sundelin et al. [21] found increased Fas in lingual SCC with no expression in healthy oral mucosa. There is, thus far, a single paper relating to FasL expression in oral carcinogenesis by Lora et al. [22]. They reported that FasL expression correlated negatively with degree of dierentiation and apoptosis in OSCCs after studying eight samples of normal oral mucosa and 19 OSCCs using immunohistochemistry and the TUNEL methods [22]. We, therefore, studied the expression of Fas and FasL in various stages of oral carcinogenesis using a previously described immunohistochemical method [23]. 2. Materials and methods 2.1. Samples A total of 64 samples were studied; 57 samples were obtained from patients attending the Stomatological Hospital, West China University of Medical Sciences, Chengdu, Sichuan, People's Republic of China, during 1994±97. They were all archival specimens graded according to the criteria of an international collaborative group on oral white lesions and the World Health Organization on oral cancers [24,25]. These included nine cases with hyperkeratosis, 24 oral premalignant lesions (OPLs, nine of mild, moderate, and six of severe dysplastic lesions), 24 OSCCs. None of the patients had radiotherapy or chemotherapy or other interventional palliative or therapeutic measures prior to sampling. Control samples of oral mucosa (5 mm diameter) were obtained from seven individuals undergoing plastic surgery, all of whom gave informed consent for the use of their tissue in the experimental procedures. The specimens were obtained and blocked, ®xed with 10% buered formalin and embedded in paran
549
using conventional histopathological techniques. Sections (3-mm) were prepared and evaluated after staining with haematoxylin and eosin. Their histopathologic grades were recon®rmed according to the standard criteria [24,25] by two experienced histopathologists, in a double-blind manner. 2.2. Antibodies The Fas and FasL rabbit polyclonal antibody was commercially purchased from Santa Cruz Biotechnology, Inc., CA, USA. According to the manufacturer, the Fas antibody was raised in rabbits immunised with a peptide sequence mapping the carboxy terminus of the Fas precursor of human origin, and was speci®c for the latter. The FasL antibody was also raised in rabbits immunised with a peptide sequence mapping the amino terminus of FasL of rat origin, and reacted with FasL from mouse, rat and humans. 2.3. Staining procedure Immunohistochemistry for detecting both Fas and FasL was performed as described previously, with a slight modi®cation, using a labelled streptavidin±biotin (LSAB) immunohistochemistry assay [23]. In brief, 3mm-thick sections of paran-embedded tissues were cut, mounted onto slides coated with 5% APES (3aminopropyltriethoxysilane, Sigma), dried overnight at 56 C, dewaxed in xylene and rehydrated through descending graded alcohols to phosphate-buered saline (PBS, pH 7.4). For antigen retrieval, slides were immersed in 10 mM sodium citrate buer (pH 6.0) and boiled twice for 5 min in a microwave oven (800 W). The liquid level was adjusted with addition of distilled water after the ®rst 5 min. Slides were then allowed to cool down in the citrate buer and washed thoroughly in three changes of PBS. In order to inhibit endogenous peroxidase activity, the slides were treated for 10 min with 3% (v/v) H2O2, 18% methanol (v/v) in PBS. Then, they were placed in a humidi®ed chamber and covered with 10% normal porcine serum (Dako, Denmark) for 10 min at 37 C to block non-speci®c proteinbinding sites. After decanting the excess serum, slides were incubated with primary antibody (diluted 1:400 for Fas and 1:100 for FasL) for 60 min, at 37 C for Fas and, overnight, at 4 C for FasL. Slides were subsequently incubated with prediluted biotinylated antibody against rabbit immunoglobulins (Dako), and streptavidin± horseradish peroxidase conjugate for 30 min at 37 C. After washing, peroxidase activity was detected using 3,30 -diaminobenzidine (Dako) as chromogen with H2O2 as substrate. The sections were counterstained with Harris haematoxylin, dehydrated, cleared in xylene, and mounted in DePeX (DaKo).
550
Qianming Chen et al. / Oral Oncology 35 (1999) 548±553
2.4. Interpretation of results The cells in the test specimens, which demonstrated granular brown staining, were considered as positive. The expression of Fas and FasL was estimated semiquantitatively as follows: ÿ, no expression; + (mild), <5% positive cells; ++ (moderate), 6±25% positive cells; +++ (intense), 26±50% positive cells; or ++++ (very intense), >50% positive cells [16]. To minimize inter-examiner variation, each slide was evaluated by two experienced histopathologists and the specimens with more than 10% deviation in the recorded staining intensity index were reassessed using a multi-headed microscope to achieve consensus. 2.5. Controls Positive controls were included in each assay and comprised formalin-®xed, paran-embedded sections of normal human lymph nodes. Blank controls were fabricated for each specimen by the omission of the primary antibody, which was replaced with PBS. Substitution of the primary antibody with antibody-free serum was also used as a replacement control. 2.6. Statistical analysis Kruskal±Wallis test was utilized to evaluate the differences in the staining intensity of Fas or FasL in each group. Dierences at 5% level were considered statistically signi®cant. All statistical evaluations were performed using the SPSS version 6.0 package (SPSS Inc., Chicago, USA).
Table 1 Expression of Fas in oral carcinogenesisa Staining
N
HK MiD MoD SD WOSCC MOSCC POSCC
++++ 0 (Very severe)
0
0
0
0
0
0
0
+++ (Severe)
1
2
0
0
0
0
0
0
++ (Moderate)
5
6
4
2
1
0
0
0
+ (Mild)
1
1
4
7
3
6
2
0
ÿ (Negative)
0
0
1
0
2
6
6
4
Total number 7 of specimens
9
9
9
6
12
8
4
a Staining: Fas and FasL expression were semi-quantitatively evaluated as follows: ÿ, no expression; +, <5% positive cells; ++, 5± 25%; +++, 26±50% positive cells; ++++, >50% positive cells. N, normal controls; HK, hyperkeratosis; MiD, mild dysplasia; MoD, moderate dysplasia; SD, severe dysplasia; WOSCC, well-dierentiated oral squamous cell carcinomas (OSCCs); MOSCC, moderately dierentiated OSCCs; POSCC, poorly dierentiated OSCCs.
tended to be either equivalent to or greater than that of the normal mucosa (Fig. 1b). In the invasive OSCC group, the degree of expression of Fas antigen decreased signi®cantly when compared with the controls ( p<0.05) and seemed to relate well to the degree of tumour dierentiation. For example, in the 12 well-dierentiated tumors, six samples demonstrated a mild degree of antigen-positive cells (<5%, according to the grading system) particularly in cell nests (Fig. 1c). No Fas-positive cells were observed in the epithelial cells in poorly dierentiated OSCCs.
3. Results
3.2. FasL staining in oral epithelia
3.1. Fas staining in oral epithelia
Generally, FasL expression was mainly seen in the cytoplasm, with little or no expression on the cell membrane. The overall results of FasL expression in healthy and diseased tissues are shown in Table 2. FasL could be detected in all normal oral mucosal tissues, particularly localized in the basal cells, with a weak reaction in the prickle cell layers (Fig. 2a). In hyperkeratotic tissues and OPLs, the staining pattern of FasL was similar to normal controls, although the number of FasL-expressing cells was signi®cantly higher ( p<0.05) (Fig. 2b). All 24 specimens from OSCCs demonstrated increased staining intensity as well as cell numbers with FasL ( p<0.05; Fig. 2c) when compared with both normal controls and OPLs.
The Fas antigen appeared as granular material essentially on the cell membrane of the epithelium, although some positive reaction was visualised in the cytoplasm. The results of Fas expression during various stages of oral carcinogenesis are shown in Table 1. Immunohistochemically, Fas was expressed in all seven normal oral mucosal specimens. This was clearly evident in more dierentiated layers such as the granular and the upper part of the prickle layer, but not in the basal cell layers (Fig. 1a). Varying degrees of Fas expression were seen in the diseased tissues. The staining pattern of all nine hyperkeratotic lesions was similar to the normal oral mucosa. Quantitatively, the number of positive staining cells in the three OPL groups appeared to be lower when compared with the normal controls. This was particularly true in the moderately and severely dysplastic tissues ( p<0.05). However, the intensity of staining
3.3. Fas and FasL in the lymphocytic in®ltrate Both Fas and FasL could be detected in the lymphocytic in®ltrate of the lamina propria of the normal
Qianming Chen et al. / Oral Oncology 35 (1999) 548±553
551
Table 2 Expression of FasL in oral carcinogenesisa Staining
N
HK MiD MoD SD WOSCC MOSCC POSCC
++++ 0 (Very severe)
0
0
1
3
8
4
4
+++ (Severe)
0
1
4
6
1
3
3
0
++ (Moderate)
1
5
1
2
2
1
1
0
+ (Mild)
6
2
3
0
0
0
0
0
ÿ (Negative)
0
1
1
0
0
0
0
0
Total number 7 of specimens
9
9
9
6
12
8
4
a Staining: Fas or FasL expression were semi-quantitatively evaluated as follows: ÿ, no expression; +, <5% positive cells; ++, 5± 25%; +++, 26±50% positive cells; ++++, >50% positive cells. N, normal controls; HK, hyperkeratosis; MiD, mild dysplasia; MoD, moderate dysplasia; SD, severe dysplasia; WOSCC, well-dierentiated oral squamous cell carcinomas (OSCCs); MOSCC, moderately dierentiated OSCCs; POSCC, poorly dierentiated OSCCs.
the normal mucosa. However, the number of antigenbearing cells tended to increase proportionally with the degree of lymphocytic in®ltrate. 3.4. The controls In each case, positive controls (i.e. human lymph node tissues) for both Fas and FasL, were consistently positive, whereas negative controls (sections without primary antibody and sham rabbit serum for Fas and FasL) were consistently negative. 4. Discussion
Fig. 1. Fas immunohistochemistry in oral carcinogenesis. (a) Fas expression in normal oral mucosa, predominantly in the cells of the granular layer and the upper part of the prickle layer; no Fas antigen was detected in the basal layers but some reaction was noted in the in®ltrating lymphocytes (arrow). (b) Some evidence of Fas staining on cell surface in moderate dysplasia. (c) A well-dierentiated tumour, where the Fas reaction appeared mainly in cell nests.
mucosa (Figs. 1a and 2a) as well as OPLs (Fig. 2a and b), and in all the tumours (Fig. 2c). The degree of expression of both these antigens in the lymphocytic in®ltrate of OPLs and OSCCs was similar to that of
Our results clearly indicate that Fas antigen is expressed to some extent in the well-dierentiated layers of the normal oral mucosa as well as in cells of welldierentiated SCCs. The latter was present mostly in the centre of tumour nests. These ®ndings seem to suggest that Fas expression is related to the degree of dierentiation of the squamous epithelium, not only in normal but also in carcinomatous tissues. These results concur with those of Yoshioka et al. [17] and Gratsa et al. [16] who reported down-regulation of Fas expression in oral mucosa and oesophageal mucosa, respectively. However, Sundelin et al. [21] could not detect Fas expression in normal oral mucosa. We also clearly demonstrate Fas positivity particularly in cell membranes, a ®nding consistent with previous observations that Fas is a type II membrane protein [1,2] where the carboxy terminus is membranic.
552
Qianming Chen et al. / Oral Oncology 35 (1999) 548±553
Fig. 2. FasL immunohistochemistry in oral carcinogenesis. (a). Fas L staining cells were detected in the normal oral mucosa, mainly in the basal cells, with a weak reaction in the prickle cell layer. The expression of FasL was also seen in the in®ltrating lymphocytes in the lamina propria (arrow). (b) Moderate dysplasia: the cell number and the intensity of FasL expression is increased compared with normal controls. (c) Dysplastic epithelium from an oral squamous cell carcinoma (OSCC) demonstrating increased FasL staining intensity compared with normal controls as well as the oral premalignant lesions, with simultaneous FasL expression in in®ltrating lymphocytes (arrow).
The Fas antibody, which we used in the study, was raised with a peptide sequence mapping at the carboxy terminus of the Fas. Fas expression was signi®cantly down-regulated in the oral premalignant lesions. Indeed, in the majority of OSCCs investigated such expression was completely lost or limited to a small population of neoplastic cells. This suggests that down-regulation of Fas expression may be an early event in oral carcinogenesis in an attempt to escape immune surveillance and destruction by activated cytotoxic T cells [9±15]. Similar ®ndings have been reported by Strand et al. [9], Grate et al. [16], Yoshioka et al. [17] in hepatocellular, oesophageal, and oral cancers, respectively. However, these reports are inconsistent with the ®ndings of Sundelin et al. [21], who found increased Fas expression in SCC of the tongue. We also demonstrate here that FasL is constitutively expressed in the normal oral mucosa, predominantly in the basal and suprabasal layers. The biological signi®cance of FasL expression in such a graded fashion in normal oral epithelium needs to be further investigated. As there is a degree of lymphocytic in®ltrate in normal oral mucosa [26], it is temping to speculate that FasL expression may contribute to the protection of basal cells from apoptotic cell death induced by lymphocytes that are present in the region. FasL positivity was detected mostly in the cellular cytoplasm both in the normal tissues and the diseased epithelia, thus agreeing with previous studies [1,2]. This result is not surprising as the amino terminus in this family of proteins is cytoplasmic, and the FasL antibody we used in this study was raised in rabbits immunised with a peptide sequence mapping at the amino terminus. Our results imply that up-regulation of FasL may be common in OSCCs as this phenomenon was readily detectable in oral premalignant lesions. In general, the degree of FasL expression concurred with the degrees of dysplasia, suggesting this to be an early event in the natural history of oral carcinogenesis. This is consistent with the recent report by Lora et al. [22] who correlated the immunohistochemical expression of FasL with the dierentiation of OSCCs and found that FasL expression was highly expressed in poorly dierentiated lesions and in some moderately dierentiated lesions. We have also demonstrated that the lymphocytic in®ltrate in OSCCs expressed both Fas and FasL whereas in a previous study we reported that the number of the in®ltrating lymphocytes signi®cantly decreased in such lesions [26]. These may be related events. Recent reports on hepatocellular carcinomas [15], melanomas [10], lung carcinomas [11], glioblastomas [12,13], colon carcinomas [14], and oesophageal carcinomas [16] clearly suggest that the FasL expressed in carcinoma cells is functional. Such functional FasL in tumour cells is likely to defend the cells against attack from the cytotoxic T cells or
Qianming Chen et al. / Oral Oncology 35 (1999) 548±553
other eector cells of the immune system by triggering an apoptotic response in lymphocytes [10]. Therefore, we propose that up-regulation of FasL is a defensive strategy of neoplastic cells of the oral epithelium attempting to escape immune rejection. In summary, constitutive expression of Fas appears to correlate with the degree of dierentiation of the normal oral mucosa. The majority of invasive OSCCs demonstrated down-regulation of Fas-expressing cells and simultaneous in®ltration by FasL-expressing lymphocytes, suggesting this to be a mechanism by which these cells evade Fas-mediated cell death. Furthermore, the majority of invasive OSCCs show up-regulation of FasL when compared with the premalignant lesions, implying that tumour cells may counteract the immune system by killing Fas-expressing cytotoxic T cells. Thus, upregulation of FasL and down-regulation of Fas appear to be simultaneous events in oral carcinogenesis. These results, taken together, indicate that the Fas/FasL system may serve as a useful clinical tool in the future as a prognostic biomarker in oral premalignant lesions. Acknowledgements This work was supported in part by a grant from the National Nature Science Foundation in China (No. 39300149), a grant from the Ministry of Public Health in China (No. 9504125), and a grant from the Committee on Research and Conference Grants of the University of Hong Kong, Hong Kong (No. 10201937/ 30713/08011/301/01). References [1] Nagato S. Apoptosis by death factor. Cell 1997;88:355±65. [2] Nagato S, Golstein P. The Fas death factor. Science 1995; 267:1449±56. [3] Leithauser F, Dhein J, Mechtersheimer G, et al. Constitutive and induced expression of APO-1, a new member of the nerve growth factor/tumor necrosis factor receptor superfamily, in normal and neoplastic cells. Laboratory Investigation 1993;69:415±29. [4] Watanabe-Fukunaga R, Brannan CI, et al. The cDNA structure, expression and chromosomal assignment of the mouse Fas antigen. Journal of Immunology 1992;148:1274±9. [5] Frech LE, Hahne M, Viard I, et al. Fas and Fas ligand in embryos and adult mice: ligand expression in several immune-privileged tissues and coexpression in adult tissues characterized by apoptotic cell turnover. Journal of Cell Biology 1996;133:335±43. [6] Suda T, Takahashi T, Golstein P, Nagata S. Molecular cloning and expression of the Fas ligand, a novel member of the tumor necrosis factor family. Cell 1993;75:1169±78. [7] Hahne M, Renno T, Schroeter M, et al. Activated B cells express functional fas ligand. European Journal of Immunology 1996; 26:721±4. [8] Loughran TPJ. Clonal diseases of large granular lymphocytes. Blood 1993;82:1±14.
553
[9] Strand S, Hofmann WJ, Hug H, et al. Lymphocyte apoptosis induced by CD95 (APO-1/Fas) ligand-expressing tumor cells: a mechanism of immune evasion? Nature Medicine 1996;2: 1361±6. [10] Hahne M, Rimoldi D, Schroter M, et al. Melanoma cell expression of Fas (Apo-1/CD95) ligand: implications for tumor immune escape. Science 1996;274:1363±6. [11] Niehans GA, Brunner T, Frizelle SP, et al. Human lung carcinomas express Fas ligand. Cancer Research 1997;57:1007±12. [12] Saas P, Walker PR, Hahne M, et al. Fas ligand expression by astrocytoma in vivo: maintaining immune privilege in the brain? Journal of Clinical Investigation 1997;99:1173±8. [13] Gratas C, Tohma Y, Van Meir EG, et al. Fas ligand expression in glioblastoma cell lines and primary astrocytic brain tumors. Brain Pathology 1997;7:863±9. [14] O'Connell J, O'Sullivan GC, Collins JK, Shanahan F. The Fascounterattack: Fas-mediated T cell killing by colon cancer cells expressing Fas ligand. Journal of Experimental Medicine 1996; 184:1075±82. [15] Shiraki K, Tsuji N, Shioda T, Isselbacher KJ, Takahashi H. Expression of Fas ligand in liver metastases of human colonic adenocarcinomas. Proceedings of the National Academy of Science, USA 1997;94:6420±5. [16] Gratas C, Tohma Y, Barnas C, et al. Up-regulation of Fas (APO1/CD95) ligand and down-regulation of Fas expression in human esophageal cancer. Cancer Research 1998;58:2057±62. [17] Yoshioka C, Muraki Y, Fukuda J, Haneji T, Kobayashi N. Identi®cation of the Fas antigen in human gingiva. Journal of Dental Research 1996;75:1353±7. [18] Muraki Y, Yoshioka C, Fukuda J, Haneji T, Kobayashi N. Immunohistochemical detection of Fas antigen in oral epithelia. Journal of Oral Pathology and Medicine 1997;26:57±62. [19] Dekker NP, Lozada-Nur F, Lagenaur LA, MacPhail LA, Bloom CY, Regezi JA. Apoptosis-associated markers in oral lichen planus. Journal Oral Pathology and Medicine 1997; 26:170±5. [20] Chrysomali E, Dekker NP, Regezi J. Apoptosis in oral erythema multiforme. Oral Surgery Oral Medine Oral Pathology Oral Radiology Endodontics 1997;83:272±80. [21] Sundelin K, Jadner M, Norberg-Spaak L, Davidsson A, Hellquist HB. Metallothionein and Fas (CD95) are expressed in squamous cell carcinoma of the tongue. European Journal of Cancer 1997;33:1860±4. [22] Lora LL, Vintermyr OK, Johannessen AC, Liavaag PG, Jonsson R. Suppression of Fas receptor and negative correlation of Fas ligand with dierentiation and apoptosis in oral squamous cell carcinoma. Journal of Oral Pathology and Medicine 1999;28: 82±7. [23] Chen Q, Luo G, Li B, Samaranayake LP. Expression of p16 and cdk4 in oral premalignant lesions and oral squamous cell carcinomas: a semi-quantitative immunohistochemical study. Journal of Oral Pathology and Medicine 1999;28:158±64. [24] Axell T, Pindborg JJ, Smith CJ, Waal Van Der I. Oral white lesions with special reference to precancerous and tobacco-related lesions: conclusions of an international symposium held in Uppsala, Sweden, May 18±21, 1994. Journal of Oral Pathology and Medicine 1996;25:49±54. [25] Pindborg JJ, Reichart PA, Smith CJ, Waal van der I. Histological Typing of Cancer and precancer of the Oral Mucosa, 2nd Edition. World Health Organization, Springer, 1997, p. 11±13. [26] Nie M, Chen Q, Li B, Zhen G, Luo G, Zhen X. A quantitatively pathological study on the lymphocytic in®ltrate in oral carcinogenesis. Clinical Journal of Stomatology 1997;13:213±5 (in Chinese).